Data Lake / Data Warehouse Modernization · Marketing Personalization
Capillary Technologies needed to process and analyze massive loyalty and marketing datasets in time to deliver real-time personalized campaigns, but its legacy data warehouse was slowing performance and increasing infrastructure and maintenance costs.
Reported impact
Capillary increased marketing campaign conversion rates by 600 percent, reduced compute and infrastructure costs by 35 percent, cut inference queries from 150 seconds to 115 seconds, saved 180 hours of MLOps overhead, and got new models into production faster to improve retention performance.
Capillary rebuilt its personalization stack on the Databricks Data + AI Platform using a lakehouse architecture instead of its overloaded legacy data warehouse. Loyalty, transaction, and registration data from more than 400 brands flow into a unified environment that combines lake-scale storage with warehouse performance, letting the team process more than 13 billion transaction records per year without prior infrastructure drag. Capillary uses high-performance data pipelines and managed clusters on Databricks to prepare analytics for campaign execution, while MLflow simplifies the model lifecycle for targeting and retention models. Databricks SQL gives internal users visual access to results, and Delta Lake provides the shared storage foundation for analytics and machine learning. By removing manual infrastructure and MLOps work, the implementation lets Capillary launch personalized campaigns faster, improve retention, and control resource usage as client onboarding grows.
Loyalty, transaction, and registration data from Capillary's client programs land in the Databricks lakehouse, pipelines and managed clusters prepare them for analytics and machine learning, MLflow helps operationalize models, and Databricks SQL exposes insights for teams running and optimizing personalized campaigns.
From an operations perspective, the automation and sheer performance of the lakehouse allowed them to reduce inference queries from 150 seconds to 115 seconds — reducing customer resource usage and costs by 35%.
Using Capillary’s algorithm on Databricks Data + AI Platform, the conversion rate of the campaign increased by 600%.
With Databricks Data + AI Platform, they are able to streamline cluster management and build highly performant and reliable data pipelines to feed their analytics and machine learning needs.
With MLflow, we have been able to build an ML workflow with ease, saving us 180 hours of MLOps overhead.
With over 100 global loyalty programs in flight at any given time, Capillary processes more than 13 billion transaction records per year while recording 187 million new customer registrations during the same time period.
With over 13 billion customer records to analyze, Capillary’s legacy infrastructure buckled under the weight of their clients’ data sets — slowing performance and creating operational overhead.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Databricks Sample Sep 22, 2025 Sovrn To operate at this scale, Sovrn needed to turn siloed Commerce and Ad Exchange data into actionable insights. Previously, fragmented systems, costly infrastructure and complex governance slowed innovation and increased… 30% Increase in publisher yield with smarter audience targeting and traffic shaping AI/BI · Delta Lake · Delta Sharing Databricks Sample Aug 22, 2025 ServiceNow However, the path to growth was mired in inefficiencies such as siloed data, disconnected systems and time-consuming manual workflows. 1.5x Boost in deal velocity Delta Lake · Lakeflow Jobs · Unity Catalog Databricks Sample Jun 13, 2023 InMobi InMobi needed to replace an expensive multicloud data warehouse because rising data volumes, siloed teams, and limited openness were hurting reporting performance and product agility. 15% Databricks Data + AI Platform · Databricks SQL · MLflow 